Transcription
The new Hermes agent just released, and it completely destroys cloud code for both building things and for using it to automate tasks. It's more secure, has more scalable AI, and it's an always-on agent that you could chat with through Telegram or WhatsApp. By the end of this video, you're going to know exactly how to set it up and five insane use cases that you could begin to use it for today.
Now, that brand new way to be able to deploy Hermes in seconds is going to be using Abacus AI because now when you actually come into Abacus AI, you're going to see this right here: "Deploy Hermes in seconds." Then, once you actually click on this, you're going to be brought to this screen right here where it says, "Welcome to Hermes agent. Connect your AI model provider in order to begin to actually use this." And we can see right here that we can always set up or manage our providers later via the provider icon in the tab bar. And by the way, if you don't have a COD API key, you don't have an OpenAI Codex API key, you could use Abacus AI in order to actually power this right now if you wanted to.
And then this is actually going to walk you through exactly how you do this. So, this is going to walk you through exactly how to sign in to Anthropic, be able to go through the console, be able to actually create your API key, and it shows you exactly what to do with it once it's actually done. Then, you're simply just going to come over here into COD, walking through exactly what it wanted you to do in those instructions. And then you're just simply going to come back over here, paste in your API key, and then you're going to see that this is now connected and ready to go. So, we could see that COD has been connected. It's going to be using our Anthropic API key, which is then going to be saved to our Hermes computer, and we've been automatically switched to COD Opus 4.7. And if we want to, we could change models at any time via the terminal.
Now that this is set up, this is exactly what this is going to look like. Right here, we can see all the available skills that it has for all these different things like research, social media, GitHub, gaming, email, DevOps, creative, autonomous AI agents, and so much more. And we simply chat with it through the terminal right here. In addition to that, we can see on the left-hand side over here that we have terminal, we have dashboard, we have desktop. And I want to dive in a little deeper into what all this actually does. So, when we actually come over here and set this up, here's what it looks like. We could actually start our computer right here and begin to actually use our agent. And once it's connected, we can actually come over here into the dashboard. We'll be able to see our sessions, our analytics, our cron jobs, which are going to be scheduled tasks, our skills, examples, config. We have different keys right here. And this is pretty cool where we can actually go through and see everything. For example, hooking this up to Telegram or different API servers. And what I like about this a lot is this actually allows us to see analytics around a daily breakdown, different models, different skills that were used, and all the things that this tool is actually doing right here. And we can see that it has built-in for 7-day, 30-day, 90-day, and we could refresh it.
On top of that, if we click on cron right here, this is where you're actually going to be able to go through and schedule things out. For example, we can name all of our different automations. We could put in the prompt that we want to happen. We could add in when it is actually going to happen and where it's going to be delivered. For example, we could have it delivered on Telegram. We could have it delivered on WhatsApp or directly actually in our local instance. And in addition to that, we could go through right here all the different tool sets that this has. For example, we could give it access to Airtable. We could give it access to cloud code, to cloud design, and more than 79 different things with more than 23 different tool sets in a bunch of different areas: general, autonomous agents, creative, data science, DevOps, email, gaming, GitHub, MCP, media, note-taking, productivity, and so much more.
Then, the other thing that's actually incredibly helpful here is this also has documentation. So, we have user stories and use cases, getting started using it, different features, different messaging platforms, integrations, and I really like how well this is documented here. And another thing that's pretty awesome is you could actually just chat with it in order to determine what all it can do. And then another thing I wanted to show you here is actually getting this set up with Telegram. You're simply just going to come over here and you're going to chat with this, for example, "Connect with my Telegram account and help me create an Hermes-powered personalized bot there that can remember my conversations, tasks, learn from them, and have more personalized context over time." Then, from here, this is literally going to walk you through the exact instructions on how you get this set up. It's going to be two options: we have option A or option B right here. And then it says, "After you actually do that, I can configure all of these different things." And what's even great is it gives you security notes so that you're not giving this access to anything that you'll regret later.
And now that this is actually connected, we could see everything that this can do. This can now manage files locally, automate tasks, use Telegram-friendly outputs, generate creative content, analyze images, send messages, connect to platforms, remember useful preferences, and this actually gives us a list of things. For example, we could ask this to connect with our GitHub account and check for any open PRs and email the respective owners in order to actually close them. And then check out what this does. This actually then goes through, uses all these skills, uses to-dos, uses the terminal, and then actually gives us a response here, which is that it needs a little bit more access. But we can see that this now allows you to do all of this stuff from your phone 24/7.
Now, before I get into the rest of the use cases that are only going to keep getting crazier and crazier, I wanted to remind you that you could go to the pin comment below and get started with the new Hermes agent on the Abacus AI supercomputer in seconds. And not only does this tool let you deploy this new agent, it also allows you to build viral mobile apps, access every LLM from one location, and so much more. So, what are you waiting for? Go to the pin comment below and get started with it today so you could deploy it literally in seconds.
Okay. So, we can see that now that we're actually set up here, what we're going to notice is all these different tools that it has, all the different skills that it has. Again, I could come over here into the dashboard if I wanted to, which is pretty awesome. If we actually come over here and type in something like slash, what you were going to notice is that we have a bunch of different options here. We could start a new session. We could refresh our session. We could actually go through the conversation history. We could save a current conversation. We could have it actually redo the last message, which is pretty awesome, and we could get it to do a bunch of other things like title conversations, branch, fork, compress, which is really cool that it has all of these different capabilities.
But the first use case that I wanted to show you is how you could actually use this to create way better prompts for anything that you're doing. For example, I'm going to say to this right here that I want to build a social media AI tool that helps people write content for X, but I don't know the perfect prompt for it. Can you design the perfect prompt? And then we can actually use this in order to get this done. And now we can see that this is now going through and this is doing exactly that. And really, what's most powerful here is that it is actually using skills that it comes preloaded with in order to actually help you create this right here. So, we can see that here's a battle-tested, production-ready system prompt for X Twitter content writing assistant. It's designed to be dropped directly into an API tool, an API call, or an agent configuration. And that is literally, it goes through the system prompt, goes through the identity, goes through the voice and tone adapters, goes through the default tone, goes through the output rules, the workflow, and this in seconds gave us a perfect prompt. It gives us threads, replies, quote posts, examples of good and bad, how to actually use this prompt, and then we can actually get this to use this from inside of here. So, essentially, what I'm really showing you right here is an all-for-one use case. So, this is a use case that could be used for anything. Doing something like this: "I am trying to accomplish XYZ. Can you create me the perfect prompt for this?" Then, it's actually going to create that prompt. And then what we could actually say to this is, "Let's use this prompt around a post about Hermes agent. I want three of them." And then what this is actually going to do is this is going to go through and this is going to write me three perfect X posts that I could actually post. And I can now get this to do this on a daily basis by literally just coming over here and saying something like, "Can we create a cron job so that you do this on a daily basis at 9:00 a.m. Eastern?" And then this actually goes through and this has some questions that we need to clarify. How do we want it to do it? Do we want it to save to a file for the same topic every day? Do we want to be able to save a file only but rotate topics daily? Do we want to send to a messaging platform, for example, Telegram? Or do we want to do both or something else? And then we can reply to this, tell it exactly what we wanted to do. And boom, this will actually set up that automation or cron job for us. And now when we actually come over into our cron jobs right here, we could see that this is scheduled. This is ready to go. And this is created.
Now, the next thing that I really like this for is doing types of technical research and competitive breakdowns. For example, we could come over here and say something like this: "I am thinking about breaking into the baseball apparel niche with an e-commerce brand. Can you perform a full analysis on how I should break into this space? My competitors are companies like BL101.com and Baseballism. Give me a complete business plan in order to compete with them and reasoning of why it will actually work." And we can see that this again goes through and this actually has all of these different skills. It has web extract, it has web search, and it's going to go through and it is going to pull all these different trends. It's going to perform all this technical analysis. And what this then allows us to do is come up with a business plan incredibly quickly. And it now has the context that we are trying to do this. And we could set up a bunch of cron jobs around this. And we can actually get this to help us do this. And if we actually look at this, it's insane how many things this is actually going through and doing here. And again, what's incredible about this is that we don't actually have to go through and download these skills, give it access to all this. It came with all of this preloaded.
Now, we can see that this actually has gone through and this created this. So, we have an executive summary right here, which is pretty incredible. We have our competitor analysis, and this actually goes through and pulls in all this different information. We have the same exact thing from baseball. Here, we have a market gap identified right here. So, this actually goes through where they would recommend actually breaking into this space based on the different competitors and things that they're missing. We have brand positioning right here. Product strategy. We have expansion. And this literally breaks down like a 12% strategy that we should use right here, or sorry, a 12-month strategy that we should use. Supply chain and operations, goes through marketing and customer analysis. Goes through community and retention. Goes through different financial predictions right here. Revenue targets, unit economics, and this is incredible. It also goes through why this will actually work. We have a 90-day launch roadmap right here. We have risks and mitigations. Then we have final thoughts. And again, it performed all of this literally in like a minute.
Now, the next use case I wanted to show you right here is actually going to be a bunch of different things you can do inside of the messaging platform right here. So, we can see that we could message us on Telegram, on Discord, on Slack, on WhatsApp, on Signal, on email, and all of these different tools right here. In addition to that, if we actually scroll down a bit right here, we'll be able to see all the different integrations that we could give this access to through MCP AI providers, through API servers, through provider routing, through fallback providers, through credential pools, and so much more. In addition to that, if we actually come over here to user stories and use cases, this literally goes through a bunch of different use cases. And we could see tons here. They have use cases for personal assistant, for integrations, for creative, for trading and markets. For example, this one right here is actually one of my favorites. We could actually click into this and we could see exactly what this is right here. It walks through the prompt and it walks through the story and the use case. And I would strongly suggest that you go through all of these different stories, all of these different use cases because there are going to be so many of these that are going to be relevant to what you are trying to do, and you will then be able to backtrack it.
And what you could actually do with all these different stories and use cases is come over here into Hermes and you could say something like, "I want to recreate this use case that I saw in your documentation. Walk me through exactly how to do it or just set it up yourself." Then, we're actually going to come over here. We are going to grab this right here and we are going to get everything that we need right here. We will copy this. We'll come back over here. We're going to paste this here. And then this is actually going to go through. This is going to review it. This is pulling up the Poly Market skill, all of these other skills. And now this is actually going to walk through exactly how this gets done, how to actually do it. And it's crazy because all we're doing is actually finding a use case, then saying to this, "Hey, go off and figure out exactly how to do this." And now look, it's actually doing this right here, which is incredible because all you have to do now is be like a curator of use cases, skills, and stories. And the agent itself will figure out exactly what to do. And I want you to look at all that this is doing because again, this is absolutely insane. And now I want you to look at what this did. This actually went through and created the four-layer trading framework right here that looks at an order book analysis, goes through recent trade flow, goes through price momentum, goes through market metadata right here. This connected to Poly Markets API and all of this, and it is going to run this every single time right here, which is absolutely insane. And now we could get this to allow it once, allow it for this session. We could add it to a permanent allow list. Now, this will actually go through and do this for us. And again, they have all of these different use cases that you can actually go through. You could look at them and you could just drop them in here, copy and paste. Boom. This will go through and this will actually spin up exactly how to do it, copy how to do it, or figure out how to do it, which is insane because now you have a 24/7 AI employee that can connect to anything, that can schedule anything, that can learn how to do anything. And this is insane that you could access all of this right here.
So, what are you waiting for? If you want to begin automating your business 24/7, you need to go to the pin comment below and get started with Hermes Agent on the Abacus AI supercomputer in seconds. In addition to that, if you enjoyed this video, I would strongly suggest you check out this video right here that walks you through a bunch of recent updates from Google IO. I'll see you over there.